Accuracy indicators

Definition and presentation of accuracy indicators

Type of data collection	Accuracy indicators	Definition and presentation
A: Census	Response rate	achieved number⁽¹⁾/ frame population number Presented as %
	Coverage rate	total value of production of the respondent units/total value of production of the frame population Presented as %
	Coefficient of Variation (CV) only if response rate <70% ⁽²⁾	where: ô(Ŷ) is the estimate standard deviation of Ŷ Ŷ is the estimated total value per fleet segment of the variable e.g. total energy costs ⁽³⁾ Presented as absolute value (0.2 rather than 20%)
B: Probability Sample survey C: Non-Probability Sample survey	Achieved sampling rate	achieved sample number⁽⁴⁾/frame population number Presented as %
	Coverage rate	total value of production of the respondent units/total value of production of the frame population Presented as %
	Response rate	achieved sample number⁽⁴⁾/planned sample number⁽⁵⁾ Presented as %
	Coefficient of Variation (CV)	where: ô(Ŷ) is the estimated standard deviation of Ŷ Ŷ is the estimate of the total ⁽³⁾ Presented as absolute value (0.2 rather than 20%)

(1) Achieved Number is the number of respondents who supplied data in response to the census

(2) CV is also required for census which achieves a low response rate (<70%) as this must be treated as if it were a Non-Probability Sample survey

(3) The estimated total is the final estimate for each variable and each fleet segment, according to appendix VI of DCF (e.g. estimated total energy costs, estimated total crew costs, per fleet segment).

(4) Achieved sample Number is the number of respondents that supply data (and not, for instance, the number of questionnaires sent out, or number of companies contacted)

(5) Planned sample Number is the number of units to be contacted for the survey (even though you may not expect all of them to respond and supply data)

reference: STECF report (11-19)